Fei-Fei Li Helped Create AI, Now She Feels the Responsibility

0:00So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300 ,000 can use AI to fill their HR questions, resolving 94 % of common questions. Not noise. Proof of how we can help companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business. When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies, that isn't always easy.

0:38Risk can touch multiple parts of an organization at the same time, often in ways that aren't immediately obvious. It might involve property, liability, or cyber. It could stem from regulatory requirements or challenges tied to a specific industry or the scale of an operation. At that level, managing risk becomes an ongoing discipline, not a one-time decision. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. That means working with companies to identify where they're exposed, decide what matters most, and put practical standards in place so risk is managed as part of day-to-day operations.

1:14And when losses do happen, the Hartford can pair that risk control work with insurance coverage grounded in underwriting, risk engineering, and claims experience developed over time. Learn more at thehartford.com slash risk mitigation. From game day crowds to memorable meals, Genius by Global Payments keeps your kitchen and floor perfectly in sync. Real-time menus, seamless updates, big league reliability for any business. That's genius.

1:45Bloomberg Audio Studios. Podcasts, radio, news. I was always a curious kid. I wasn't curious about nightclubs or other things. I was an avid lover of science. Fei Fei Li, tech CEO, Stanford professor and godmother of AI. Are you conscious of the power that you have? I'm conscious of my responsibility. I understand I'm one of the people who brought this technology to our world. The agency is not the machines, it's ours. From Bloomberg Weekend, this is The Michelle Hussain Show. I'm Michelle Hussain.

2:38Have you ever thought about where AI came from? The origin story of tools like ChatGPT, Claude or Gemini, which are part of many people's everyday lives. I don't mean the principle of training large language models on data, but the long, often lonely work of the scientists who believed years ago that you could teach machines to be intelligent. Well, today's guest has lived that story because she was at the forefront of unlocking this world for us all. And she has a deeply resonant personal story. She was 15 when she came to the US from China with her parents. She had little English, life was hard, money incredibly tight.

3:27But she excelled at school and set herself audacious problems to solve. Her name is Fei-Fei Li. And by the end of our conversation, I was asking myself, why do most people know the name Sam Altman and not hers? She's the one whose breakthrough was critical. No wonder they call her the godmother of AI. We ended up talking about everything from her own childhood to, yes, whether AI will do away with us humans and what we should teach the kids of today. Her answer on that might just surprise you. But I hope you enjoy our conversation as much as I did. Dr. Feifei Lee, thank you for coming on the show.

4:16I'd love to start with this remarkable period for your industry. It's three years since ChatGPT was released to the public. And since then, there have been new vehicles, new apps, huge amounts of investment flowing towards the industry. How does this moment feel to you? Well, first of all, Michelle, thank you for inviting me to this show. I'm very excited. It's a good question because AI is not new to me. I've been in this field for 25 years. I live and breathe it every day since the beginning of my career. Yet this moment is still as daunting and almost surreal to me in terms of its massive, profound impact.

5:01This is a civilizational technology. And it's surreal personally to me because I'm part of the group of scientists that made this happen. And clearly I did not expect it'll be this massive. When was the moment it changed? Because I know you've talked about the years when it was like an AI winter. Yeah. And what you're describing now, how the moment feels extraordinary even to you, is it because of the pace of developments or is it because the world has woken up to it and therefore turned the spotlight on people like you? I think it's intertwined, right? But for me, to define this as a civilizational technology is not about spotlight.

5:44It's not even about how powerful it is. It's about how many people it impacts. Everyone's life, work, well-being, future will somehow be touched by or impacted by AI. In bad ways as well as good? Well, technology is a double-edged sword, right? Yes, I think both ways, because since the dawn of human civilization, we create tools we call technologies. And these tools are meant in general for doing good things. But along the way, we might intentionally use it in the wrong way or we might have unintended consequences. And strands of this, I know, will emerge in different ways over the course of this conversation.

6:31But you said the word power. And I'm struck by the fact that the power of this technology is in the hands of a very small number of companies, most of them American. How does that sit with you? You're right. The major tech companies, through their massively reaching products, they are impacting globally our society the most. I would personally like to see this technology being much more democratized. I would like to see no matter who builds or holds the profound impact of this technology, do it in a responsible way. And I also do believe every individual in this era should feel they have the agency to impact this technology.

7:23We'll talk a bit more in a moment, I think, about the democratization and how that might be achieved. But you, of course, in terms of companies in this field, you're part of that because you are a tech CEO as well as an academic. In fact, I think your very young company, little more than a year old, is reportedly already worth a billion dollars. Yes. I'm co-founder, CEO of World Labs, and we are building the next frontier of AI, which is spatial intelligence, which people don't hear too much about today because we're all about large language models. Yet I believe spatial intelligence is as critical and complementary to language intelligence.

8:04Releasing this idea of virtual worlds, which again we'll dig into. But before we do that, I want you to take us back again to the fact that you've seen the whole trajectory of this industry. You've been in it for 25 years. I know that your first academic love was physics. What was it in the life or work of the physicist you most admire that made you think beyond that particular field? I grew up in a, it's not a small town, but it's a less well-known city in China. And I come from a small family. So you could say life was small in a sense. It was in the 80s, my childhood, which was fairly simple and isolated.

8:46You were an only child? I was the only child. Off on the way in China under the one-child policy. And my family in general is just small. Physics is almost the opposite. It's vast. It's audacious. The imagination is unbounded. You look up in the sky, you can ponder about the beginning of the universe. You look at a piece of a snowflake, you can zoom into the molecular structure of the matters. You think about time. You think about magnetic field, you think about nuclear. It takes my imagination to places that you can never be in this world. And what really fundamentally fascinates me to this day about physics is not be afraid of asking the most boldest, audacious question about our physical world, of our universe, of where we come from.

9:42But your audacious question, I think, was about what is intelligence? Yes. And I think it's rooted in my love for physics and my training in physics that I look at each physicist I admire, look at their audacious question, right? From Newton to Maxwell to Schrodinger to Einstein, my favorite physicist. And I wanted to find my own audacious question. And somewhere in the middle of college, my audacious question shifted from physical matters to intelligence. What is it? How does it come about? And most fascinatingly, how do we build intelligent machines? And that became my quest, my North Star. And that's a quantum leap because from machines that were doing calculations and computations, you're really talking about machines that learn and that are constantly learning.

10:41I like you use the physics pun, the quantum leap. Yeah, crossing over into ordinary parlance. But I think, didn't you have this light bulb moment? As you're thinking about intelligence, you realize that a critical part of that is our ability to recognize objects. The fact that around us right now, there are multiple objects, we know what they are. The ability that humans have about recognizing objects in the world is foundational. So I decided that was my first law start. My PhD dissertation is to build machine algorithms to recognize as many objects as possible. I think the key question is how, right?

11:23How do you teach the machines? And what I found really interesting about your background is that you are reading very widely and there are two key breakthroughs that make your ultimate breakthrough possible. And one is that you start thinking and learning about what psychologists and linguists are saying that is related to your field. Well, that's the beauty of doing science at the forefront. Because it's new, no one knows how to do it. As an AI scientist, you look for possible answers, and it's pretty natural to look at human brain and human mind and try to understand or be inspired by what humans can do.

12:04And one of the inspiration I got in my early days of studying or trying to unlock this visual intelligence problem is look at how our visual semantic space is structured. There are so many tens and thousands and millions of objects in the world. How are they organized? Are they organized by alphabets or are they organized by size or colors? And you're asking that because you have to understand how our brains organize in order to have something to teach the computers. That's one way to think about it. And I come upon this linguistic work called WordNet. WordNet is a way to organize semantic concepts, not visual, just semantic or words in a particular taxonomy.

12:54Give me an example. An example is in a dictionary, an apple and an appliance are very close together. But in real life, an apple and a pear are much closer to each other. And the apple and pear both belong to fruits, whereas an appliance belong to a whole different family of objects. And I came upon this taxonomy called WordNet that describes this organization of objects. And I made a connection in my head is that two things is very clear. One is that this could be the way visual concepts are organized because apples and pears are much more connected than apples and wash machines, for example.

13:42But also even more importantly is the scale, is that if you look at the number of objects described by language, you realize how vast it is, how big it is. And that was a particular epiphany for me. We as intelligent animals, as humans, we experience the world with massive amount of data. And we need to endow machines with that ability. Hence the big data sets that you needed. And actually, it's worth noting that at that time, I think this is early part of the century, the idea of big data. 2006, 2007. Doesn't really exist. No, big data was not. This phrase did not even exist. The kind of scientific data sets we were playing with were tiny.

14:34How tiny? For example, in images, most of the graduate students in my era were playing with data sets of four or six or at most 20 object classes. In each class, there were a couple of hundred examples at most, and that's how tiny it is. Whereas fast forward three years later, after we created ImageNet, it had 22 ,000 classes of objects and 15 million annotated images. And ImageNet was a huge breakthrough. It's the reason that you've been called the godmother of AI. I'd love to understand more about what it is about you that enabled you to make these connections and see things that other scientists didn't.

15:21And one that immediately comes to mind is that you acquired English as a second language after you moved to the United States. Is there something in that that led to what you're describing? Michelle, that's a great question. I've been asked this question and my true answer is I don't know. Human creativity is still such a mystery. People talk about AI doing everything and I disagree. There's so much about the mystery of human mind that we don't know. So I can only conjecture that a combination of my interests and my experiences led to this, including the curiosity and the hunger for audacious problem definition.

16:05I find myself, I'm not afraid of asking crazy questions in science, not afraid of seeking solutions that's out of box. And maybe my appreciation for the linguistic link with vision might be accentuated by my own journey of learning a different language, but I don't know the answer. What was it like coming to the United States as a teenager? Because I think that's a particularly difficult stage in life to move and to have to make new friends, even if you weren't battling a language barrier. That was hard. I think being an immigrant, I came to this country when I was 15 in New Jersey, Precipita in New Jersey.

16:55None of us, my parents, nor I spoke much English at all. I was young enough to learn more quickly. It was very hard for my parents, and we were not financially very well at all. my parents were doing cashier jobs and I was doing Chinese restaurant jobs. Eventually, by the time I entered college, also my mom's health was not so good. So my family and I decided we have to run this little dry cleaner shop in New Jersey to make some money to survive. Right. And you got involved yourself. I joked I was the CEO. I ran the dry cleaner shop for seven years. From what age? I was 18, 19 to the middle of my graduate school.

17:43So all the way through college and graduate school. Yes, seven years. Even when you're at a distance, you're running your parents' dry cleaning business. Yes, because I was the one who spoke English. So I take all the customer phone calls. I deal with the billing. I deal with inspections or the business. What did it teach you? Resilience. As a scientist, you have to be resilient because science is a nonlinear journey. Nobody has one conjecture and all the solutions in front of you. You have to go through such a challenge to arrive and find an answer. And as an immigrant, you learn to be resilient.

18:22Were your parents pushing you because they clearly wanted a better life for you? That's why they left China, for the United States. So how much of this was coming from them and how much from your own sense of responsibility to them? To their credit, they did not push me much. They're not tiger parents in today's language. I think part of it is they're just trying to survive, to be honest. And especially my mom, she is an intellectual at heart. She loves reading. But the combination of a tough, survival-driven immigration life plus her health issues, she was not pushing me at all. As a teenager, I kind of had no choice.

19:09I either have to make it or not make it, and the stakes are pretty high, so I was pretty self-motivated. I was just curious. I was always a curious kid, and my curiosity had an outlet, which was science. And that really grounded me. I wasn't curious about nightclubs or other things. I was an avid lover of science. But also an avid reader that from your early childhood, you're putting away the children's books and turning towards the classics. The grown-up books. My mom liked reading and she probably thought that you're talking about when I was nine, ten. At that time, she probably did have an influence, more influence.

19:56She thought I was precocious enough that just read some grown-up books. You also had a teacher who was really important in your life. Tell me about him. I excelled in math, and I liked math. And I befriended the head teacher, Mr. Bob Sabella, and the math teacher. We became friends through the mutual love of science fiction reading. Which you'd read in Chinese. At that time, I was reading in Chinese. Eventually, I started reading English. He's a remarkable person because he probably saw in me this desire to learn. So he went out of his way to create opportunity for me to continue to excel in math.

20:47I remembered I placed out of the highest curriculum of math. And so there was no more courses for me. He would use his lunch hours to create a one-on-one class. for me to do. Now that I'm a grown-up, I know he was not paid extra. It was really out of a teacher's love and sense of responsibility. And he really became such an important person in my life. Is he still alive, Mr. Sabella? Mr. Sabella passed away when I was an assistant professor at Stanford. But his family, his two sons, his wife and I, I think they're my New Jersey family at this point. You used the word love about what he did for you.

21:33And I wonder if they also, he and his family, introduced you to American society. Were they your first friends and an entrance to the whole world of America beyond your school? Absolutely. They introduced me to the quintessential middle-class American family. They live in a suburban house. They have two lovely kids. Of course, they're all married and have the grandchildren generation now. But it was a great window for me to know the society, to be grounded, to have friends and have a teacher who cared.

22:15So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300 ,000 can use AI to fill their HR questions, resolving 94 % of common questions. Not noise, proof of how we can help companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business. IBM. This coffee shop? Running smooth thanks to Genius. From global payments, instant transactions, effortless inventory, and synchronized operations.

22:56Big league reliability for any business. That's Genius. Support for this show comes from Public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip. Manually sweeping idle cash. Putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S &P 500. Or, if my cash balance goes above$20 ,000, move the excess into my direct index. You approve of the workflow and your agent handles the rest.

23:33Monitoring the market, watching for your conditions, and executing your strategies exactly as defined. An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API. Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., member FINRA and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures.

24:24Do you think you could have had the career that you have had in China? Because now there are significant advances happening in AI in China. I don't think I'm able to answer this question because I think life is so serendipitous, right? The journey would be very different. And in a way, we could have simulated all possibilities. But what is timeless or what is invariant for anyone is the sense of curiosity, the pursuit of North Stars. So if I were me, I will still be doing AI somehow, I believe. Do you still feel connected to China? It's part of my heritage. I feel very lucky that my career has been in America, higher education, and then Silicon Valley, then also in and out of industry, and being in tech.

25:21The combination of all these ingredients is very global. The environment my family right now is in, which is Stanford, San Francisco, Silicon Valley, is very international. So I feel very connected. And the discipline I'm in, which is AI, is so horizontal. It touches people everywhere that I do feel much more like a global citizen at this point. And of course, it is a global industry. but there are some really striking advances in China, not least the number of patents and the number of AI public papers coming out, the deep seek moment earlier this year. As you look ahead in this century, do you think China will catch up with the US in the way that it has in other fields like manufacturing?

26:07I do think China is a powerhouse in AI. In this moment, most people would recognize the two leading countries in AI are China and US. The excitement, the energy, and also, frankly, the ambition of many regions and many countries in the world of wanting to have a role in AI, wanting to catch up or even come ahead in certain areas of AI, that is pretty universal. And your own next frontier, spatial intelligence. Tell me what you mean when you use the word spatial intelligence. What are you working on right now? Yeah, so spatial intelligence is the ability for AI, or frankly, any intelligence to understand, perceive, reason, and interact, and also create spaces, worlds.

27:02It comes from a continuation of visual intelligence. The first half of my career around the image that time was trying to solve the fundamental problem of understanding what we are seeing. And that was a very important problem. But it's not enough because that's a very passive act. It's just receiving information and being able to understand this is a cup, this is a beautiful lady, this is a microphone. But if you look at evolution, you look at human intelligence, perception is profoundly linked to action. We see because we move. We move, therefore we need to see better. And how do you create that connection has a lot to do with space.

27:51Because you need to see the 3D space. You need to understand how things move. You need to understand when I touch this cup, how do I mathematically organize my fingers so that it creates a space that would allow me to grab this cup? All this intricacy is centered around this capability of spatial intelligence. And I've looked on your website and I've seen the preview that you've released, Marble. you can see essentially a virtual world which one goes into and from room to room and one door opens and you go from one place to another but I'm not sure how you use it is it essentially to you a tool for training AI it's a different way to train AI rather than for example meta saying this is in metaverse this is a world you can go into and spend time in as a human being right so let's just be clear with the definition, marble is a frontier model.

28:53It's a frontier world model. What it does is not just to see and go into a world. What's really remarkable is it generates a 3D world at a simple prompt. A prompt could be, give me a modern-looking kitchen. Or the prompt could be, here's a picture of a modern looking kitchen, make it a 3D world. The ability to create a 3D world is a fundamental ability. It's fundamental to humans and I hope one day it's fundamental to AI. If you are a designer or architect, you can use this 3D world to ideate, to design. if you are a game developer, you can use it to obtain these 3D worlds so that you can design games.

29:47If you want to do robotic simulation, these worlds will become very useful as training data for robots. If you want to create immersive educational experiences in AR, VR, this model would help you to do that. Interesting. I'm imagining girls in Afghanistan, maybe you could do virtual classrooms in a very challenged place. Yes, or I'm imagining, for example, how do you explain to an eight-year-old what is a cell? One day, we'll create a world that's inside of a cell, and then the student can walk into that cell and understand the nucleus, the enzymes, the membranes. So you can use this for so many possibilities.

30:35So that's your next frontier. Can we, and I'm conscious yours is a very big, complex industry, but there are some immediate pressing issues. And I wonder if I could put a selection of those to you for you to give us an instinctive or even a nutshell response on how you see them. For example, and you'll have heard this many times before, number one, is AI going to destroy jobs, large numbers of jobs? Mm-hmm. Technologies do change the landscape of labor. A technology as impactful and profound as AI will have a profound impact in jobs. Which is happening already. Which is happening already. Salesforce said 50 % of their customer support roles are going to go because of AI.